Abstract: Due to technologic and societal shifts, it seems clear that patients will be much more involved in their care going forward. Health care providers and decision-makers should consider these shifting cultural and technologic dynamics in their care delivery strategies. There is much discussion needed about how providers can involve patients in their own care so that it remains safe, appropriate, and adequate. The creation of evidence-based pathways for use by patients would be the natural next step to help patients find the best, most-evidence based information rather than potentially unvetted information they find online. We discuss how providers and clinical pathway developers can evolve with these technologic and societal changes for the benefit of all stakeholders.
Historically, because of an almost complete asymmetry in knowledge and impenetrable medical jargon, medical decisions about patients’ health have been almost completely made by doctors. While well-intentioned physicians might have taken time to explain a decision and its consequences to a patient, the reality is that only a few patients would have a firm understanding of their treatment options and the relevant pros and cons. But a number of new trends are converging to enable a more informed and active patient group and general public. Many believe1-4 there is a revolution on its way, as envisaged by Eric Topol in his famous book, The Patient Will See You Now: The Future of Medicine Is in Your Hands.5
Society has changed as the internet has transformed the ease of accessing information; all doctors are familiar with the Google phenomenon leading to self-diagnoses and an increasingly curious and interested patient population that no longer takes the physician’s advice as necessarily being the best. The very different standards of care in different regions (eg, depending on whether a patient lives in rural vs urban area, distanced from a specialty center) have been well documented as have the very high rates of medical errors occurring. Indeed, prior research has estimated that medical errors may be associated with over 400,000 deaths each year in the United States.6 Patients are increasingly aware of these dynamics and are searching out the best possible treatments, even if it means considerable travel.
Simultaneously, entrepreneurs looking to create new tech opportunities have found a large market for consumer-facing apps delivering a profusion of wellness solutions relating to everything from exercise to nutrition. These apps are directly engaging a growing online community of health-conscious, often young, consumers in managing their own wellness. Now, even more medically orientated apps delivering heavyweight medical advice—often providing relatively easy access to real physicians online—are beginning to add a more medically focused dimension, as we will discuss below.
Following the increased ease of access to information through the internet and the development of patient-facing technology, the third axis contributing to the shift to a more informed and active patient group is the emergence of value-based care. As provider organizations around the world come under increasing financial and competitive demands, combined with the influences above, there is a continuing pressure to drive expertise and health management into the community and frequently into the patients’ own homes, reducing lengths of stay and admissions to acute care. Of course, the cheapest care, where it can be efficiently managed, is patient self-care. In relation to this demand for more value in care, the importance of patient-centered care and patient-related outcomes have become increasingly important.
There is much discussion needed about how providers can involve patients in their own care so that it remains safe, appropriate, and adequate. As patients are increasingly utilizing technology as a means to educate themselves in medical care—looking up their diseases, medical research, treatment options—the creation of evidence-based pathways for use by patients would be the natural next step to help patients find the best, most evidence-based information rather than potentially unvetted information they find online. In order to deploy evidence-based pathways knowledge to patients online, a number of technical capabilities will need to be successfully addressed.
Pathways in Executable Formats
First, there has to be a capability to digitize the clinical knowledge embodied in a pathway. Some may think that this is merely providing access to electronic versions of clinical guidelines online. However, in reality, this does not optimize this process any better than handing out generic leaflets about a particular disease to a patient, as a patient is unlikely to have the skill set to relate the relevant information to their own condition.
The more sophisticated strategy would be to provide a patient-user-friendly pathways digital platform that can be integrated with individual patient data to automatically generate diagnostic and treatment recommendations. This, at the moment, is more easily achieved in diagnosis, where decision trees, triage, machine learning, and predictive analytics can be directly brought to bear, rather than in treatment, where the vast variation in individual patients and the relative lack of defined probabilistic advice can make it difficult to apply such modalities comprehensively and broadly to highly sophisticated decision support.
An example of diagnostic machine learning technology is Google Deep Mind, which has been deployed at hospitals in London to identify patients at high risk of suffering acute kidney injury and to help categorize ophthalmic scan results to aid in diagnosis of retinal disease.7,8
While decision-tree models are widely used in clinical decision support (CDS) technologies and available within most electronic medical record systems, these generally are limited to situation action alerts and order sets that do not function well for less linear and more advanced treatment advice. However, decision trees are deployed to great effect in, for example, most commercially available oncology pathway systems, where the contingencies and algorithmic logic are well established and linear in nature.
Only a few companies, relatively, are currently providing generic, highly sophisticated, treatment-focused CDS, Deontics, Elsevier, and IBM Watson being examples.
The take-home message here is that the technologies to digitize clinical pathways and to deploy them over the internet already exist and will improve, providing patients with the capabilities to understand the pros and cons of the specific treatment options, to avoid unnecessary treatment, and, above all, to clarify issues affecting their own safety.
Deployment to Individual Patients
Getting high-quality, evidence-based information to patients in a way that is useful to them can be difficult. Applications need to convey complex concepts in a jargon-free and bias-free way that properly addresses important health matters. This requires validation that goes beyond a simple test of whether or not patients may or may not like a particular user interface. With the growing profusion of patient-facing care administration and remote consultation, there are some apps available to the connected patient that may hold insight for future patient pathways.
Examples of these patient-facing apps include Babylon Health, Doctor Care Anywhere, Push Doctor, NowGP, Dialogue, YourMD, and Sensely. There are now a large number of such companies that all, to greater or lesser extents, help patients connect with the right care for them in a convenient way, and often offer some automated diagnostic and data capture capabilities. The addition of sophisticated treatment advice will further add to the capabilities of such apps to empower patients. In the field of mental health, it is well recognized that patient-facing computerized behavioral therapy produced by companies such as Ieso Digital Health can work very effectively and can be more effective than direct patient/physician interactions.9
These applications build on the experience of the last 20 years of telemedicine, where clinical expertise (which may be centralized) can be deployed over large geographies. Telemedicine was, and still is in practice, largely physician-to-physician, with experts in elite institutions helping less expert physicians decide diagnoses and management plans for individual patients. Telemedicine was pioneered by companies such as Worldcare, which specializes in connecting top medical experts for second opinions on care, opening up the expertise of elite US hospitals to patients all around the world.10 Physician-to-physician telemedicine techniques are also increasingly becoming standard in regions with significant rural patient populations who may find it difficult to visit a hospital, eg, in Texas, the Canadian provinces, and others.
In addition, there are a number of companies providing digital communities for patients, where patients can go for information about their disease and to share experiences. These can also collect data to enable research. Examples of these communities include PatientsLikeMe11 and Raremark.12 These offer the opportunity for patients (“members”) to compare experiences and services, and they provide relevant information regarding specific disease states to members. In the case of Raremark, this is focused on rare diseases only (where patients may be very geographically distant around the world).
The common clinical drawback with all such remote systems is the difficulty posed to the physician of not being able to directly examine the patient physically, and the use of these technologies must always be adjusted to take this into account. Nonetheless, companies are evolving all the time as they learn from experience regarding how best to render complicated medical information to the layman and how to deliver this information to individual patients.
It is extremely important to keep in mind that patients come from very different educational backgrounds with varying capabilities to assimilate and understand information. In addition, each patient will have a different level of interest in being involved in decisions about their own care. Somehow patient facing systems need to be able to accommodate all of these factors. Indeed, the effectiveness of a user interface may become an important aspect of regulation of digital health technologies in the future.
Validation, Regulation, and Security
For the most part, current patient-facing digital solutions do not actually include specific individualized, actionable, diagnostic or treatment advice for a patient with a serious condition, thus generally these solutions are not regulated. Where they are regulated, they are principally labeled as “devices.” Regulators internationally, in general, are coming from the starting point of traditional medtech perspectives relating to the oversight of diagnostic and interventional products, where risk is regulated on the basis of levels of physical intervention and risk of injury, or of specificity and sensitivity in diagnosis and accuracy in measurement of a parameter. However, these are not always easily applicable to modern digital technologies where risk may be introduced through exertion of, sometimes vague and ill-defined, influence over clinical decision making with no physical risk being incurred or measurement being made. The result is that in many cases, regulatory status is something of a grey area.
In general, perhaps the most reliable safety measure on patient-facing pathways would be that they would always require a physician to sign off on any significant actions. However, with these developing, more sophisticated software solutions, we are not that far away from being able to put some self-management into the hands of the patient. This is desirable where it can be deemed safe, as it could potentially decrease health care costs significantly and increase the capability to manage patients in the community; however, then the question becomes, “Who decides what is safe for patients with no specialist knowledge?” As the capability of modern digital technologies continues to develop at pace, there is likely a significant role for the regulators in the future, to make sure that systems are safe. However, there is also a question as to whether the regulators are well positioned to do this.
Data security is more straight forward, being clearly applicable anywhere where patient data are captured. The recent furor over the use of individual data on Facebook has highlighted the potential for misuse, and data considerations are already a major consideration for all agencies dealing with patient data.
Shared Decision Making: The First Step
For many of the reasons discussed above, it is probably too early to put clinical pathways directly into patients’ hands, although that proposition will no doubt develop. The natural first step is enhanced shared decision-making.
For example, companies such as Deontics have the functionality, typically via existing electronic health records or patient portal user interfaces, to allow patients to see what their individualized treatment options are and what the justifications are for each option in their specific case; what the prognostics are for each option for them individually; and how their preferences affect their treatment options. This would enable a patient to be fully informed in a way that a Google search could never accomplish. It also allows the patient to intervene where a physician may be about to commit a medical error, is advising unnecessary treatment, or is advising treatment for which the evidence is equivocal.
In a recent small study that included patients with stage II colorectal cancer, wherein patients were given direct online access to such personalized information and the evidence for adjunctive chemotherapy is equivocal, only 1 of 12 patients opted to have chemotherapy.13 While there was no control arm to the study, it seems likely that more than one patient would have proceeded to have the expensive chemotherapy in the absence of access to such information.
Shared decision-making, enabled in this way, is an attractive proposition not only from the point of view of improved patient experience and patient safety, but also from the point of view of improved physician experience and improved consent, as consultations with better informed patients would be faster and more satisfactory, and there would be a comprehensive digital record of decision-making and the underpinning evidence and rationale.
Lingering Questions
Will clinical pathways be put into the hands of patients in the future? The answer seems to be almost inevitably “yes,” and quite possibly on a much shorter timescale than expected, as the pace of technological change continues to accelerate.
At present, we await clear evidence on the pros and cons of greater patient involvement in clinical decision-making. However, the benefits that we might anticipate include increased patient safety, reduced unnecessary care, and improved patient experience. At the same time, it seems very possible that effective involvement of patients could simultaneously improve physician experience, resulting from improved communication and clinical efficiency with better outcomes achieved at lower cost. In part, the latter might result from patients taking over greater self-management in appropriate circumstances, contributing to the drive to push more care away from expensive acute centers and avoiding unnecessary travel. Also, usage of acute resources might be more efficient, with reduced inappropriate admissions, reduced lengths of stay, reduced error, enhanced and better documented consent, reduced readmissions, and reduced medico-legal costs. Where these effects may seem desirable in the developed world, they may be all the more so in the developing world where levels of trust by patients in providers can be very low.
In terms of potential dangers, there is the “garbage in/garbage out” phenomenon where poor digital solutions may give patients distorted understandings of the pathways. Further, processes may become disrupted and more frustrating with physicians having to take more time to explain and disabuse patients of false, possibly harmful, impressions gained from poor digital solutions or, for example, where integration into existing workflows is only partially complete and therefore disruptive.
Conclusion
The goal of giving patients better understanding and control of their own care seems desirable and could well improve patient safety, and it has great potential to reduce costs and improve patient and physician experiences. However, if not handled well, the converse may very well come true. What seems to be clear is that the technologic elements or foundations to put pathways into patients’ hands do, to a large extent, exist already; the cultural, regulatory, and commercial forces that will advance these tools are already in place.
References
1. Drucker Mueller L. The diagnostic revolution: how patient-centered care is driving innovation. Cardinal Health. https://www.cardinalhealth.com/en/essential-insights/the-diagnostic-revolution-driven-by-patient-centered-care.html. Published April 21, 2017. Accessed August 19, 2018.
2. Richards T, Montori VM, Godlee F, Lapsley P, Paul D. Let the patient revolution begin. BMJ. 2013;346:f2614.
3 Tailor K. The Patient Revolution: How Big Data and Analytics Are Transforming the Health Care Experience. Hoboken, NJ: John Wiley & Sons, Inc; 2016.
4. Godlee F. Towards the patient revolution. BMJ. 2014;348:g1209.
5. Topol E. The Patient Will See You Now: Th Future of Medicine Is in Your Hands. New York, NY: Basic Books; 2015.
6. James JT. A new, evidence-based estimate of patient harms associated with hospital care. J Patient Safety. 2013;9(3):122-128.
7. Suleyman M. Working with the NHS to build lifesaving technology [blog]. DeepMind Technologies. https://deepmind.com/blog/working-nhs-build-lifesaving-technology/. Published November 22, 2016. Accessed August 16, 2018.
8. DeepMind Technologies. DeepMind and Moorfields eye hospital NHS Foundation Trust. https://deepmind.com/applied/deepmind-health/working-partners/health-research-tomorrow/moorfields-eye-hospital-nhs-foundation-trust/. Accessed August 16, 2018.
9. Kessler D, Lewis G, Kaur S, et al. Therapist-delivered Internet psychotherapy for depression in primary care: a randomised controlled trial. Lancet. 2009;374(9690):628-634.
10. Worldcare. Worldcare International, Inc. https://www.worldcare.com/. Accessed August 16, 2018.
11. PatientsLikeMe. https://www.patientslikeme.com/. Accessed August 16, 2018.
12. Raremark-About. Raremark. https://about-raremark.com/. Accessed August 16, 2018.
13. Miles A, Chronakis I, Fox J, Mayer A. Use of a computerised decision aid (DA) to inform the decision process on adjuvant chemotherapy in patients with stage II colorectal cancer: development and preliminary evaluation. BMJ Open. 2017;7(3):e012935.
